Self-Adaptive SRv6-INT-Driven System Adjustment in Runtime for Reliable Service Function Chaining

Self-adaptation of service function chains (SFCs) has been considered as an important attribute to ensure the resource-efficiency and reliability of network function virtualization (NFV) systems. In this work, we leverage the idea of integrating segment routing over IPv6 (SRv6) and in-band network telemetry (INT) seamlessly to realize SRv6-INT and explore the mutual benefits of SRv6 and INT for achieving self-adaptive SFC deployment. Specifically, we design and experimentally demonstrate a self-adaptive SRv6-INT-driven SFC deployment system that orchestrates network and IT resources timely to adapt to bursty traffic and network changes. We first enhance our previous design of SRv6-INT to better use it for self-adaptive SFC deployment, and then propose an IT resource management technique for Kubernetes (K8s) to accomplish resource allocation and contention resolution without offline virtual network function (vNF) profiling. Next, a closed-loop system is designed to manage SFCs in both the local and global ways. As for the local way, we let servers make local decisions based on the INT data encoded in packets to scale the vNFs running on them vertically. The global way involves the control plane, which oversees the SFC deployment in the whole network to change the number and placement of vNFs and the traffic routing through them. Finally, we prototype our proposal with commodity servers and hardware PDP switches based on Tofino ASICs, and experimentally demonstrate its effectiveness.